I’ve had nothing but issues with NC instant upload, and stopped using it. It’s error prone and needs constant hand holding for no good reason. It didn’t handle taking a picture and then deleting it instantly very well (and will throw your notifications at you for this, often more than 1). When you have limited connectivity it will utterly confuse itself and ask you to resolve conflicts for 100 files for no reason, when it could just checksum server and client files and notice they are all the exact same. Also when set to only upload on WiFi, and not being connected to WiFi it often still spams notifications that the “upload failed”, despite not being supposed to upload anything. And btw. it could upload files just fine, they failed only because upload on mobile is disabled!

It’s a nightmare. Commonly also referred to as a cluster fuck.

As far as analogies go, is pretty far off. It doesn’t hold even for basic behaviors of the two cases, let alone complex ones. A better analogy would be that you buy a (small) car that always happens to come with an included, free trailer for more cargo capacity. You can of course take it off and have a small car. And it’s also as magic trailer that doesn’t take up any space at all when not in use, but can also not be sold.

I’m actually not a fan, an am also using it somewhat reluctantly personally, though self hosted. I’ve had my issues with it, but an still using it because it solves some issues that are much harder to solve without it. I’m not using the contacts/calendar functionality.

But your original statement was that you couldn’t understand who would need calendar and contacts (in their file sharing app). There’s enough I object to in this statement that I wrote my comment. First of, in this context, specifically in their article/blog/whatever, it’s about nextcloud as a whole, not the fact that it can do file sharing. That’s what it evolved from, but not all it is any more, more better or worse. Secondly, it’s about an advertised alternative to O365, which includes the very common and almost universal requirement for teams (be it a company, family, …) to have events (=calendar) like schedule meetings with people (=contacts). Even if you work with just like 5 people you are probably gonna need that. There you probably want to share files, but probably more so it’s about the office functionality and collaborative, simultaneous editing of files. Obviously replacing Word, PowerPoint, Excel. And yes, Outlook (calendar,& contacts, also email).

This isn’t meant for individuals who need a few GB to store some files. It’s for teams of some description that need office like, cloud based tools.

Your analogy doesn’t make any sense, so I assume you really don’t know. So let me explain:

If you buy a 4wd, it’s always a 4wd, usually that means s relatively large vehicle. You might be able to turn it to 2wd, but it doesn’t make the car smaller. If you just needed a tiny car in all (or most cases), you can’t push a button to make it smaller. You always drive around the extra equipment to possibly make it 4wd.

Nextcloud is plugin based. Assuming this isn’t locked away on an instance like this, you can literally push a button and make that whole functionality go away everywhere. You can fully remove that ‘clutter’, if that’s of no use to you. They are offering it always, as it adds no additional effort on the hosters side: they don’t need to add gear boxes or whatever to make it have calendar & contacts. If you don’t want/need it, turn it off and it’s gone.

Who ever wanted a file sync platform that also does calendaring and contacts?

Most O365 & exchange users? If you just want file sync, this isn’t for you. If you want a collaborative (online) office suite that can also sync files, it is. It’s meant to be able to replace the whole O365 stack, which includes Outlook.

You can also just not use that part, or any part you don’t need. This is basic NC functionality that has been there for a very long time, so why shouldn’t it be part of the package?

That also means you can’t downvote just wrong information anymore. Look at YouTube, which disabled down votes and nothing got better. And their votes weren’t even public.

I think that’s a terrible idea.

I disagree with those saying that you can’t do a build for that budget, but I would suggest looking into used parts, at least for some things, to improve the result significantly.

Since your system goal doesn’t seem to be storage related, as nextcloud includes storage obviously, but typically isn’t used to house multi-terabyte data sets. So assuming you can make that work for the “future homelab projects” to with dual 500gig NVME as storage. Search for a used mITX board+CPU that can accommodate that (has the slots), and go from there. Things like CPU cooler, if not part of a possible mainboard+CPU bundle, should be selected after the case at that is the limiting factor for it. Didn’t skimp on RAM size if you can (new or used is fine, depends what you can get in your area).

With this list you’re basically done to get it up and running.

There is absolutely nothing “simple” about that. It sounds simple, but what does “someone has purchased a product” actually mean, in technical terms?

Let’s start basic, since this is a proposal about a federated system, there are instances. Who runs these and why? Does ever seller run an instance? can there be users/customers on those? if not, who runs the customer-instances? Who defines what a product is, and are products like communities? or more like posts? how do you correlate different sellers selling the same item, where a review would obviously apply to both? can you review a shop or seller? Are delivery services their own “entitty” and can you review those, too? When you purchase an item

Now without any answers to any of those question, let’s just go to the next level. Where are the reviews stored? in the instance where the item is sold (possibly owned by the shop)? or with the user? if it’s with the user, how does a webserver displaying an item find all the reviews for it? Does this differ between reviews for items and reviews of shops/sellers?

If a review is stored on the instance of the seller, he can just add an entry to the database stating “user x purchased item y”, and the review is valid. If the reviews are stored with the user, he can spin up an instance, and create a bunch of users there who can leave reviews, because he can mark sales as “valid” as the seller, no matter if there was any item and/or money exchanged.

I wrote all of this thinking about the classic sellers attempt at “creating good reviews to boost a product”, but there is the opposite threat of review-bombing (might be a competing product or seller, or you just don’t like pink shirts and decide to review-bomb those): How you protect against those has similarities, but reverses the roles essentially. Sellers are now the “target”, and reviewers the “threat”.

Aaaand this all is just about reviews, which have no monetary value. The platforms main goal would be to deal with physical items, exchanged for real money, and creating physical effects (like shipping). All those have to also be secured in a much more robust way. If a fake review or two slip through the cracks, who cares. But if just one valuable item goes missing (or is never shipped), or the payment for it, that’s immediately a problem.

Condescension was not the intention at all. The fact that you mention logistics only as a foot note is what lead me to believe you really didn’t understand, and it was just meant as an explanation. Amazon is just scale, in every aspect, and I don’t think that can be achieved with a federated approach in the physical retail world.

As for being constructive, you can be constructive by talking someone out of an idea. I really don’t believe there’s any viability in the idea, no matter how much I wish there was. I personally value my time, so I assume others do as well. I consider saving someones time incredibly constructive, but that only applies if you intend to pursue the idea to actually get somewhere “real” with it, let’s say reaching “profit” or improving participants existing profits.

You might enjoy spending your time figuring out solutions here, maybe you see it as an economic experiment or hobby project, so it’s fun no matter the outcome. I’m that case my comment really isn’t constructive in your situation, and I’m sorry.

Rest assured I didn’t comment out of malice.

But you’re saying yourself that those are “obviously bots”. It’s easy to ignore those. And just to be clear, I really did mean the reviews, and not the score (where the skew is less transparent).

Everyone leaving a review has to have an account somewhere in the federated network. This includes seeing up an instance just to use it for review bots, or fake votes on something. Obviously there’s is defederation and other mechanisms, and I’m sure there are ways to improve the situation. But the whole base setup is just inherently much harder to get into a trustworthy position. Even the common centralized sites (not just Amazon) have trouble getting it under control when they can “see” will the related data, for finding outliers and such. I’m just saying it’s an even harder proposition.

That’s what I said with “much hotter for longer”. If it’s constantly thermal throttling, that’s gonna be an issue. Of course OC’ing also will. 50°C just isn’t an issue. Also older models have CPUs that either don’t throttle at all, or do it less well/effectively.

The CPU is perfectly happy sitting at 50°C. It is slightly happier at 30, but it doesn’t actually help in any way unless you run into throttling, or run (much) hotter for longer. It’s fine.

Some might state that the CPU is probably gonna live longer, but seriously have you ever had a CPU die on you cause it was old (or even die at all, even)? Again, it’s fine.

Having something that mostly agitates the air (not even really moving it) like a low-hundreds-rpm fan would also work. As would using one of those passive heat pipe coolers that are also overkill (especially with a fan, but just leave that off), but have the same “number looks better” effect.

You are completely misunderstanding what Amazon actually does, and why it’s successful, despite being a shitty company. It’s first and foremost a logistics company. People can order “stuff” in many places, but if they order it on Amazon, they’ll get it by tomorrow if they order it before midnight. They got warehouse everywhere. They do (some) of their own final deliveries for anyone close to those, use the big logistics players for the rest (ups, DHL, …) while having massive volume and the power to dictate price that comes with that. The number of workers in the warehouses is actually minuscule for their size, it’s all automated. Huge up front cost, very low cost once it’s actually running.

Consumers go there because they can get literally anything. Again: warehouses. It’s also a market place but that only works (these days) because it’s THE place the people go. The reviews are also a massive point, and would be inherently untrustworthy in a federated version.

How would you ever get anyone to go to your federated version for shopping that sells like “some” things? Even if you manage to combine all those shops, you’d need a way to agree on what an item is called (or how to assign id numbers) so the same item from multiple sellers is grouped in the same offer, and many similar small things you take for granted it didn’t even ever see/notice on Amazon.

That feature kinda works, but it’s incredibly fragile. It has caused so many annoyances for me over the last year or so that I’m finally done with the thing. Just go with immich instead, less headache.

This actually sounds quite interesting. Is this controlled with DNS entries at the domain level somehow, or is the subdomain fixed/mandatory?

Yes exactly. I didn’t wanna name-drop them cause they are closed for new dynDNS signups. You can create an account to manage your own domain, but you currently can’t signup for their dynDNS service, unfortunately.

That being said, I would still highly recommend them for managing your own domain, if you’re looking for a place to host literally just the DNS part.

There are dyndns providers that support the DNS challenge that have free tiers. Those are sufficient, and you can even get wildcard certs for your subdomain that way. Perfectly sufficient for a homelab.

The native Android client just can’t do two way sync. Just put a text file or something into any folder (from the web or desktop). Now sync that folder to Android. Now edit it on the web/desktop, and look for the changes on Android (without actively telling it to “sync”). Then change the file on Android, these 2nd changes are never sent back to the server unless you explicitly tell it to “sync” again, manually. That’s what I mean with 2 way sync.

There are quite a few files where you just need that to work to use them properly, like the database of a password manager as a prime example. Mine can talk to Nextcloud natively, so I don’t need the client for that, but I was incredibly close to just switching to syncthing, if I didn’t have active users that use the web office integration of Nextcloud.

Nextcloud can’t do two-way sync on Android. At all. That’s like core functionality for the product IMHO and there’s a feature request open I think. When I found that out, I basically spit out my coffee. It’s fine if you just want to upload photos you take, that kinda works (but my god is it fragile).

Nextcloud is pretty good at quite a few things, including extensibility, but having some omissions in functionality that boggle the mind.

Another name, depending on the exact context, is “hairpin NAT”. Should make googling with the specific router OP has easier.

I think you overestimate how large the percentage of people leaving actually was. Wish it was, but don’t think the number actually leaving made a still visible dent in their traffic graph. surely some toric specific knowledge did leave though.